The source code to DTrace is now available for public download here, more info on a blog entry introducing the code here.
OpenSolaris Web Site Goes Up
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
w00t! A great day for Solaris, SysAdmin’s and Developers everywhere.
Finally some community direction for the best operating system in the world!
This is what we’ve all been waiting for. Good move Sun. We hope you’ll continue to make wise decisions like this one. 🙂
Great news, but ETA on the actual distribution? I’m a bit tired of getting teased. What’s the reason for delay?
There isn’t anything solid yet. For example, can you install OpenSolaris already?
Lets give Sun some time to get everything together. Maybe in a year or so they will release all the meaty Solaris open. So far, it’s only a website with some lines of code. 😉
http://www.blastwave.org/
Did you read this:
http://www.gentoo.org/news/en/gwn/20050124-newsletter.xml
R&B
Nice job. 🙂
irc.freenode.net
#opensolaris
I’m a Sun fan in general as it is uplifting to see so much media attention about OpenSolaris and in the forums. Now, my question is: is this the beginning of a new perspective on Sun by the internet commnuity or just the results of a vocal minority?
If you look at the opensolaris.org website, in the upper-righthand corner, there’s a form where they ask you to submit your email address to be informed of the Q2 2005 launch. This implies all of OpenSolaris will be available by June of this year at the latest. If it works out that way remains to be seen…
I think they want the suspense to nearly kill us all. Their PR seems to be working for Solaris… They seem to have good reason too.
It kind of is a big deal. It’s probaly one of the largest gambles a computer company has done. Giving away millions of dollars is really generous.
I guess they think that this will really put sun back on the map for open source and free software and I hope it does.
It looks like they will open source a lot of goodies in the commercial solaris version which makes me wonder… why buy it if your not a big corporation that needs legal protection & special needs? With Darwin vs. Mac OS X it’s pretty obvious..
with some big companies that invest in open source organizations…
sun’s investment isn’t a tax write-off.
It’s finally here…but there’s no membership or forums, maybe later.
Here http://milek.blogspot.com/2005/01/open-solaris-official-announcemen… you can find some info about Open Solaris Pilot members which already have built Open Solaris themselfs – so it’s happening. But a whole process takes time, so Sun decided to go with this limited launch to show its good will and that Open Solaris is actually happening. Read RoadMap on http://www.opensolaris.org to get more information on when OS will be released.
Can’t wait until June. Who could pass up such an amazing OS as Solaris 10 when it’- free! Surely not me!
BTW: I heard that just the move from Solaris 9-10 alone cost SUN over $500,000,000? That’s a lot of mony on R&D and development! I am amazed that they are offering it for free and yet they had to do something. Hope it works well for them, I know it will for me.
> This is what we’ve all been waiting for.
Not me.
I’ll wait and see. I wonder how many drivers have been written and how common hardware problems are going to be. I remember the early days of linux and trying to get sound/printers/video/… working. It is still a bit of a problem if you a cutting edge. Well see.
If it is ready for general use it will rise to the top.
It saddens me how quickly the public imaginary can be deleted. A bunch of geeks jumps around every time somebody throws them a bone without stopping to ask why they are being thrown a bone or to think whether it might be poisoned?
Ask yourself: Why is SUN really doing this? Did they suddenly become Mother Teresa and want to help the FLOSS community?
Until I hear a statement from SUN’s top brass denouncing the SCO scum, I’ll think of them as just that, scumbags willing to do whatever it takes to make a buck.
When SCO launched its suit, McNealy was all over the news claiming that Linux wasn’t safe, but Solaris was. Do you guys have any sense of history or memory?
We have replaced all of SUN’s hardware were super cheap boxes that are super easy to service and replace. None of this non sense of paying SUN an arm and a leg for replacement hardware. Try gettting a quote on the price of a replacement NIC for your server from SUN.
Not to mention that they completely dropped the ball on everyone who bought one of their Cobalt servers. They stopped providing security updates way too early leaving thousands of customers stranded with good hardware whose software was no longer being maintaned. If it hadn’t been for the NetBSD and Debian communities, most of that hardware would still be collecting dust.
So I suggest that all of you take a very cautious approach to anything coming from SUN. The open souring of Solaris is a last ditch attempt at relevance in a world that no longer cares for proprietary operating systems.
Sun, want to change the world? Release Java as open source immediately. That will buy you some credibility. Solaris, a nice toy that most people who do not have big iron will not appreciate. You are just too afraid that the new kids are learning Linux and thus will not recommend your expensive hardware. Had Sun really wanted to change the world, they would have released Open Solaris under a GPL or BSD-compatible license. Since they haven’t, this is just a ploy to gain mindshare and a last attempt at survival.
And the Sun fanboys don’t need to remind me of NFS or OpenOffice.org. We’ve heard it all before. OpenOffice.org was released to spite Microsoft and it did its job. It got SUN a 2 billion settlement.
So when can we expect the port to Pegasos? 😉
Why is it that so many people are whinging and whining that OpenSolaris will not be GPL. Since when did the GPL become the benchmark license to be compatable with?
Even RMS himself has admitted that the GPL 2.0 is not suited to the current climate in the market, although the GPL2 does address patents it has no where near the level of protection offered by the CDDL.
Why was there not this much criticism when Mozilla was released under the MPL? Because the community needed a good open source browser.
The problem is that the Linux fanatics perceive OpenSolaris as a threat. Well let’s not beat around the bush it is a very big threat to the dominance of Linux in the open source OS market just like the BSD’s have been and will continue to be.
Don’t expect the trolls to go back into their caves either, has the Gnome V KDE thing setlled down at all…of course not.
I think that anyone who is a true advocate of a competetive open source market will be delighted at the thought os Solaris becoming open source, I know I am.
Competition breeds innovation.
is what it takes to set up yet another front Microsoft must fight on. OpenSolaris is going to send Microsoft’s sales people into ultra-FUD slinging, I’m sure of it.
Why should the Linux community fear OpenSolaris/Solaris 10? If anything, this means more resources pumped into Linux devel to catch up to the functionailty of Solaris.
<QUOTE>
In addition, IBM supports the growth of Linux through its Linux Technology Center, where more than 600 engineers worldwide — including more than 150 employees focused exclusively on Linux on POWER — work full-time on Linux as part of the open source community. Currently, more than 7,500 IBM employees are involved with Linux in porting centers, research, services, development labs, and sales and marketing.
</QUOTE>
http://www.google.ca/search?q=cache:HJKZFBwRKQYJ:www-1.ibm.com/serv…
Hooray for IBM, HP, Intel, Red Hat, Novell and all the other corporate enties for their devotion to Linux.
Good luck to Sun, they’ll need it.
To quote someone above:
“Why should the Linux community fear OpenSolaris/Solaris 10? If anything, this means more resources pumped into Linux devel to catch up to the functionailty of Solaris.”
This has nothing to do with Linux playing catch up on functionality. It is about a change of practice for Linux, a change which Linus is as yet not prepared to make. A tool like DTrace would be extremely difficult to implement and maintain in the current Linux kernel, why, because they change the API’s with every bloody point release. A lot of developers are starting to voice concerns about this, let’s just see how Linus responds.
<< It is about a change of practice for Linux, a change which Linus is as yet not prepared to make.>>
I’m not sure how you can go about making assumptions as to the direction Linus wants to take the Linux kernel. Could you have predicted the pace of the 2.6 linux kernel development, if you were asked a few years ago? I highly doubt it.
<< A tool like DTrace would be extremely difficult to implement and maintain in the current Linux kernel, why, because they change the API’s with every bloody point release. >>
I’m not a kernel developer (Chem Engineer in training
) and I won’t even try to pretend I know the workings of dtrace (I started learning D about 10mins ago), so I really couldnt comment.
<< A lot of developers are starting to voice concerns about this, let’s just see how Linus responds. >>
And you know how? The attitude I’ve felt from the majority of kernel developers ( I read LKML a bit
) is that they shouldnt be burdened by the mistakes of the past.
grepping the code shows:
/usr/src/uts/common/dtrace/dtrace.c:7243-7244
* ID. (N.B.: the temptation is being successfully resisted to
* have this cmn_err() “Holy shit — we executed this code!”)
Heh.
Although OpenSolaris won’t have a community the size of the Linux community right away, development will pick up. This will force IBM, Novell and co. to pick up the pace and it will drive innovation on both sides.
MS of course won’t pick up the pace, they’ll just fling more shit….erm….FUD and claim they innovate. Since Solaris and Linux will be major competitors and will advance even faster, they both will leave MS choking on their dust. So although I’ll probably stick with Linux for the distant future, I welcome Open Solaris.
One thing I am expecting to happen, however, is someone to fork Solaris out from under Sun the second it is open sourced. I hope it doesn’t happen, as it will confirm Sun’s fears of Java forking should it be open sourced….although if IBM released their JDK as F/OSS we won’t have to worry.
I posted the last response (or might not be in a minute, depends if someone else posts or not). Anyways, the reason I think someone may fork Solaris is that a lot of F/OSS advocates don’t trust Sun. I’m surprised it hasn’t been done with OpenOffice already (well, SOT Office and now LBA Office is sort of a fork but it doesn’t stray too far from OpenOffice ever).
anybody can tell me if it is possible and LEGAL to port DTrace to linux kernel?
thanks
<<Why aren’t we ever seeing them contributing anything to the community. Alot of talk but no commitment…>>
Yeah I know your trolling but I’ve got time to kill
– 600 full-time engineers working on the Linux kernel
– Eclipse/SWT
– Cloudscape
… read more here http://www-136.ibm.com/developerworks/opensource/index.html
*fingers crossed* If the rumors are true and IBM does release an opensource JDK, they will be my GOD.
I’m not saying that where mistakes have been made they should not be rectified. What I am saying is the Linux Kernel is a constantly shifting target at the moment. If every change in the binary interface is indicative of a major design flaw some serious thought should be put into what is going wrong.
I am getting sick of unstable drivers on current Linux Kernels because the drivers never get a chance to mature before the abi’s are changed…again
<<anybody can tell me if it is possible and LEGAL to port DTrace to linux kernel? >>
Nope. The CDDL is (intentionally) incompatable with the GPL.
<<If every change in the binary interface is indicative of a major design flaw some serious thought should be put into what is going wrong.>>
So your saying:
Repeated binary breakage = major design flaw
Hmmmm….
take Windows for example. MS has kept binary compatability (applications) for more than 10yrs. Since there is no breakage, Windows must have been designed flawlessly?
I don’t see how that logic works.
Are tools really installed in the kernel? you can package a distro with many diff. licenses.. (the concept of installing desktop apps with diff licenses on windows)
please point me out if im wrong. please!
<<Are tools really installed in the kernel?>>
I thought it was after reading one of those “Meet the Architects” articles on the Sun web page. Maybe I’m wrong though, it was a couple of months ago.
Yeah if its a seperate package, then its a no brainer.
Taking the linux kernel and changing one line of source code doesn’t equal a fork. Most of your so called FOSS advocates couldn’t code their way out of a paper bag, so it’s irrelevant if they are scared because their linux religion is threatened.
grepping the code shows:
/usr/src/uts/common/dtrace/dtrace.c:7243-7244
* ID. (N.B.: the temptation is being successfully resisted to
* have this cmn_err() “Holy shit — we executed this code!”)
Heh.
Crap — I completely forgot about that comment! And here after I said that there wasn’t any cursing in comments…
it’s obviously an issue of contention..
take Windows for example. MS has kept binary compatability (applications) for more than 10yrs. Since there is no breakage, Windows must have been designed flawlessly?
I don’t see how that logic works.
Wow, talk about logic 101: if A implies B, !A does not imply !B.
breakage = design flaw
no breakage != no design flaws
There is more than one way to have a flawed operating system outside lack of binary compatibility.
It would be possible to port the userland components to Linux, keeping the license as CDDL. Reading Bryan’s blog, it seems little would have to be touched in libdtrace or the commands. You would just have to gut the libctf and libproc dependencies.
But you couldn’t copy any of the kernel code, because in order to mix it with GPL code it must be re-licensed under the GPL. Depending on how self-contained the modules are, you can get away with an entire kernel module being non-GPL, but you’ll get a nastygram about tainting your kernel; this is a contentious point amongst kernel developers and GPL advocates. It would also depend on how many hooks were needed outside the modules themselves (clearly it’s non-zero).
I tought they “controlled” every single bit of code that has ever touch
any Sys V code…
If you recall SysVR4 was codeveloped by SUN and AT&T and was basically a merger of SysVR3 and SunOS codebases. I assume that SUN thus has lot more rights to the base SysVR4 code then most Unix code liceenses. That and i recall reading that they actually completly bought out their license a couple of years ago (supposedly HP did the same)
I think they want the suspense to nearly kill us all. Their PR seems to be working for Solaris… They seem to have good reason too.
I’m not so sure. This, today we do, today we don’t, in the future we will messages sent by various Sun executives regarding Solaris gives me the message that they have no control of what direction they move their company. This makes them less credible as business partner regardless if you like opensource or not.
It kind of is a big deal. It’s probaly one of the largest gambles a computer company has done. Giving away millions of dollars is really generous.
Only if you actually could sell it to cach in that money.
In the Solaris case, Linux and IBM have largely been eating Suns lunch lately. I really can’t see that that they have much choise. Either that, or they will see IBM and others perfect Linux as an enterprise level OS, and of course IBM will not sell Sun hardware.
IBM says they have 600 developers, but on the LKML I haven’t seen close to that many projects announced by IBM or code contributed. Maybe they are working on porting IBM to power, or porting their userland applications to that.
On a side note
This is nothing but a victory for floss. Who could have imagined 10 years ago that Sun would release Solaris’s source-code to the unwashed masses, and IBM and SGI would be supporting floss.
take a look at this article
http://www.theinquirer.net/?article=20928
Linux zealots…cut Sun some slack will ya..
solaris strength’s are in the Kernel. ZFS, Dtrace, TCP/IP stack, etc.
The rest of the OS is playing catchup to linux. The tools? why the GNU tools for the most part are being ported to Solaris to improve those. Servers, interface? etc, are in general ports of standard F/OSS software. Maybe their version of X will be nice, but don’t expect much else.
As for kernel toys, Dtrace is slowly being created by kprobes(which are now dynamic if you actually look),
The big thing is ZFS, which might get a GPL port.
As for developers, there are 100 of them now on ‘Open Solaris’ Geez how many developers did Sun have working on Solaris? It’s not to hard to see where those first 100 came from. Also Sun said they would release the entire OS this month, now it’s June. Let me Guess June will turn into August?
> – 600 full-time engineers working on the Linux kernel
This absolutely can not be true. According to Linus Torvalds there are only about 200 of more or less active contributors to Linux kernel world wide and there only a few dozen who are the “regulars” contributing big chunks of code. So this 600 full time employee figure is a complete and utter bullshit blown out of proportion. Don’t listen to IBM’s baloney, I would be surprised if they have 600 developers working on anything Linux related (utilities, apps and everything).
Competition will allways be good for foss comunity.
Linux *Bsd and OpenSolaris will compete. They will evolve into something interesiting.
The main competitors [imo] will be OpenSolaris and Linux. They will be both feature complete with equvalents for every feature of the other product. [think dtrace and kprobes]
Maybe in a year we will choose between linux and opensolaris [and bsd ofcource] but neither opensolareis nor linux will ever extinct. Sun [and opensolaris comunity] (the CDDL comunity) will have to contribute to the GPL comunity. [think Looking Glass, X.org, Gnome…].
IMO when opensolaris and linux will finish catching-up with each other, they will be `mostly` the same.
Remember FOSS was always about choise, now we have one more `item` to choose, compete, evolve and maybe play catch-up with.
Excuse me for my bad english =]
FYI
http://www.blastwave.org/
Now this too: http://www.blastware.org/
Things are moving!
Alliances have been formed.
People are involved.
Hardware vendors are involved.
We have the source!
Anyone know what compilers are used to compile Solaris for x86? Intel compilers? SUN proprietary?
gcc 3.4.3 now ships (as of Solaris 10) as /usr/sfw/bin/gcc (on both SPARC and x86), and is the compiler used to compile all of amd64 right now. The x86 side (32- and 64-bit) can be compiled completely with gcc. I don’t know about the SPARC side.
gcc 3.4.3 now ships (as of Solaris 10) as /usr/sfw/bin/gcc (on both SPARC and x86), and is the compiler used to compile all of amd64 right now. The x86 side (32- and 64-bit) can be compiled completely with gcc. I don’t know about the SPARC side.
Was just wondering since SUN sells a C compiler as a seperate product It seems to me that the SUN C compiler would be the prefered way to compile any OpenSolaris source code. Perhaps i am confusing the Sparc side of things with the x86 side but the source is supposed to be the same for both platforms. This opens up a whole new realm of porting. Imagine Solaris running on Power 5 or PA-RISC…
Look at OpenVMS, check out how open it is…right, Sun being ‘Open’ is never going to happen
I don’t know why is everybody so hostile to Sun. They may have screwed a lot of things in the past, but hey, once they’ve open sourced their OS, they can’t just take it back — so what are you scared of? Everybody can customize the OS, everybody can sell support, so it’s not like you’re going to have vendor lock-in here if something goes wrong with Sun. But IMHO they have everything to succeed: excellent technologies, tons of expertise, and, what’s extremely important, stability of the codebase. Due to the stable ABI writing drivers will be easy (and it will be hardware manufacturers who will do it — that’s the proper way, after all), so I think we will eventually see more and more hardware that is supported on Open Solaris but not on, say, Linux, or supported better on Open Solaris than in other Unix-like systems. Vendor lock-in won’t be an issue here, too (unlike Windows), because the underlying OS is open source. The same goes for ISVs. And, finally, all your favorite apps and DEs are still here — after all, they are what matters, not the kernel that runs under the hood, be it Linux, Solaris or FreeBSD. So future looks bright in any case for open source and proprietary SW users and developers alike.
Everytime there’s a Sun thread, people are comparing Linux to Solaris and vice versa. Is there a particular reason why the BSDs are never compared to Solaris in these threads?
… would be buying Trolltech and offering Qt toolkit free of charge for commercial development on Open Solaris. Now that’s a sure way to attract developers. After that they should drop Gnome and fix KDE instead.
(To the Gnome lovers: sorry. Gnome is cute and simple, I like its UI philosophy, but KDE is light years ahead in terms of technology. It’s easier to fix KDE interface than to bring modern technology to Gnome.)
If you look at *BSD and Linux – they look [somewhat] the same.
The userland? mostly the same…
So in a few years we will be choosing the kernel, not the OS.
BSD will always be there…
Linux will have to evolve to compete with solaris kernel and vice verca…
So the next battle will not be an OS one, but the kernel one =]
Sorry for my baad english.